On 2017 Jan 24, GARRET STUBER commented:

• This post-publication peer review was written for a group assignment for NBIO733: Circuits and Behavior – A journal club course organized by Dr. Garret Stuber at the University of North Carolina at Chapel Hill. This critique was written by students in the course, and edited by the instructor.

The recent work by Kim, J. et al. provides excellent evidence of genetically, spatially, and functionally distinct cell types in the basolateral amygdala (BLA). Studies aimed at molecular marker discovery often turn to less precise experiments such as bulk mRNA sequencing or proteomics in heterogeneous tissues to profile cell markers. In this study the authors first identify changes in the transcriptional profiles of fos+, activated cells following controlled delivery of appetitive or aversive stimuli to correlate with transcriptional markers. The authors convincingly demonstrate that two distinct cell populations marked by the expression of unique genes (Ppp1r1b+ and Rspo2) are preferentially activated by reward-related or aversion-related stimuli.

In open discussion we discussed the use of Cartpt-Cre to represent Ppp1r1b+ cells (as opposed to using/generating a Ppp1r1b-Cre mouse). While the supplementary figures demonstrate that this Cre-driver mouse also labels several Ppp1r1b- cells (~23%), the authors demonstrate consistent functional properties of Cartpt-Cre cells across multiple behavioral and electrophysiological paradigms. This provides strong evidence that the use of this animal is justified and informative of their proposed circuit. Additional experiments to further verify the use of this animal could test positive and negative valence in the context of other sensory modalities (olfactory, gustatory, etc.) comparable to the initial c-Fos expression experiments.

An additional concept that would be interesting to explore is the relative ‘strength’ each population of neurons appears to have with respect to positive and negative valences. Behaviorally, it appears Rspo2 neurons may have a greater influence on their respective valence (Figure 4b-e) as well as having a larger antagonistic effect (Figure 5a-f). This is a difficult claim to make considering the opposing behavioral paradigms cannot be considered to have equal strength in their respective valence. However, stronger antagonistic silencing illustrated by c-Fos (Figure5g-i) and cell recordings (Figure 6a-h) bring up the possibility. Importantly, the fact that Rspo2 neurons outnumber Ppp1r1b neurons (Table1) in the BLA may contribute to this. The potential for antagonistic microcircuits through local inhibitory interneurons is an additional avenue to explore. Another factor in this circuitry is that Rspo2 cells form direct synapses with other Rspo2 cells and vice versa in to form a synchronous circuit upon stimulation. Establishing whether these cell types share connectivity with neurons of the same (or similar) identity would be informative of the dynamics of the circuit.

An important note the authors highlight is the likelihood that different cell subpopulations reside within the Rspo2+ and Ppp1r1b+ neuron groups. Further exploration of markers found in the initial microarray analysis may shed light on these subpopulations and provide insight as to how BLA cells are programmed to function. Additionally, next generation single cell RNA-sequencing techniques could provide the necessary acuity in transcriptomic profiling of these potentially heterogeneous cell types.

The diversity of projection termini from these neurons also suggest cell heterogeneity and highlight what are possibly the most interesting findings of this article. Kim, J. et al. build on previous evidence that diverging circuits and cell populations encode positive and negative valence information separately in the BLA1, 2. Here, the authors successfully label and characterize these cell populations, but note important differences in projection targets not realized in previous studies. Firstly, Kim, J. et al. found that positive valence-associated neurons project to the medial nucleus of the central amygdala (CeM), contrary to previous findings that found projections to this area are largely associated with aversive stimuli 1, 2. Secondly, Kim et al., found that both positive and negative valence BLA neurons project to the nucleus accumbens (NAc). This builds on previous by Namburi, P. et al. showed that inhibiting NAc projecting BLA neurons did not affect fear or rewarding behavior in the context of conditioned learning1. The proposed heterogeneity of NAc projecting BLA neurons described in the current paper may account for this.

To conclude, the recent findings of structurally, spatially, and functionally antagonistic neurons in the BLA provide an interesting and important avenue to further dissect circuit architecture underlying complex behaviors as well as providing a genetic entry point into better understanding these circuits.

[1] Namburi, P. et al. (2015). A circuit mechanism for differentiating positive and negative associations. Nature. 520, 675-678. [2] Beyeler, A. et al. (2016). Divergent routing of positive and negative information from the Amygdala during memory retrieval. Neuron. 90, 2, 348-361.

On 2017 Jan 24, GARRET STUBER commented:

• This post-publication peer review was written for a group assignment for NBIO733: Circuits and Behavior – A journal club course organized by Dr. Garret Stuber at the University of North Carolina at Chapel Hill. This critique was written by students in the course, and edited by the instructor.

The recent work by Kim, J. et al. provides excellent evidence of genetically, spatially, and functionally distinct cell types in the basolateral amygdala (BLA). Studies aimed at molecular marker discovery often turn to less precise experiments such as bulk mRNA sequencing or proteomics in heterogeneous tissues to profile cell markers. In this study the authors first identify changes in the transcriptional profiles of fos+, activated cells following controlled delivery of appetitive or aversive stimuli to correlate with transcriptional markers. The authors convincingly demonstrate that two distinct cell populations marked by the expression of unique genes (Ppp1r1b+ and Rspo2) are preferentially activated by reward-related or aversion-related stimuli.

In open discussion we discussed the use of Cartpt-Cre to represent Ppp1r1b+ cells (as opposed to using/generating a Ppp1r1b-Cre mouse). While the supplementary figures demonstrate that this Cre-driver mouse also labels several Ppp1r1b- cells (~23%), the authors demonstrate consistent functional properties of Cartpt-Cre cells across multiple behavioral and electrophysiological paradigms. This provides strong evidence that the use of this animal is justified and informative of their proposed circuit. Additional experiments to further verify the use of this animal could test positive and negative valence in the context of other sensory modalities (olfactory, gustatory, etc.) comparable to the initial c-Fos expression experiments.

An additional concept that would be interesting to explore is the relative ‘strength’ each population of neurons appears to have with respect to positive and negative valences. Behaviorally, it appears Rspo2 neurons may have a greater influence on their respective valence (Figure 4b-e) as well as having a larger antagonistic effect (Figure 5a-f). This is a difficult claim to make considering the opposing behavioral paradigms cannot be considered to have equal strength in their respective valence. However, stronger antagonistic silencing illustrated by c-Fos (Figure5g-i) and cell recordings (Figure 6a-h) bring up the possibility. Importantly, the fact that Rspo2 neurons outnumber Ppp1r1b neurons (Table1) in the BLA may contribute to this. The potential for antagonistic microcircuits through local inhibitory interneurons is an additional avenue to explore. Another factor in this circuitry is that Rspo2 cells form direct synapses with other Rspo2 cells and vice versa in to form a synchronous circuit upon stimulation. Establishing whether these cell types share connectivity with neurons of the same (or similar) identity would be informative of the dynamics of the circuit.

An important note the authors highlight is the likelihood that different cell subpopulations reside within the Rspo2+ and Ppp1r1b+ neuron groups. Further exploration of markers found in the initial microarray analysis may shed light on these subpopulations and provide insight as to how BLA cells are programmed to function. Additionally, next generation single cell RNA-sequencing techniques could provide the necessary acuity in transcriptomic profiling of these potentially heterogeneous cell types.

The diversity of projection termini from these neurons also suggest cell heterogeneity and highlight what are possibly the most interesting findings of this article. Kim, J. et al. build on previous evidence that diverging circuits and cell populations encode positive and negative valence information separately in the BLA1, 2. Here, the authors successfully label and characterize these cell populations, but note important differences in projection targets not realized in previous studies. Firstly, Kim, J. et al. found that positive valence-associated neurons project to the medial nucleus of the central amygdala (CeM), contrary to previous findings that found projections to this area are largely associated with aversive stimuli 1, 2. Secondly, Kim et al., found that both positive and negative valence BLA neurons project to the nucleus accumbens (NAc). This builds on previous by Namburi, P. et al. showed that inhibiting NAc projecting BLA neurons did not affect fear or rewarding behavior in the context of conditioned learning1. The proposed heterogeneity of NAc projecting BLA neurons described in the current paper may account for this.

To conclude, the recent findings of structurally, spatially, and functionally antagonistic neurons in the BLA provide an interesting and important avenue to further dissect circuit architecture underlying complex behaviors as well as providing a genetic entry point into better understanding these circuits.

[1] Namburi, P. et al. (2015). A circuit mechanism for differentiating positive and negative associations. Nature. 520, 675-678. [2] Beyeler, A. et al. (2016). Divergent routing of positive and negative information from the Amygdala during memory retrieval. Neuron. 90, 2, 348-361.